November 09, 2005


In Slate, Witold Rybczynski argues that urban sprawl isn't a uniquely American phenomenon—something spurred on by our automobile-heavy culture and single-use zoning laws. No, sprawl's something that happens in all cities at all times; it even happened in ancient Rome! "Sprawl is and always has been inherent to urbanization." Um, okay. But that doesn't mean the sort of sprawl you see in American cities is always inevitable, right? Clearly there are degrees and different types here. So what causes which sorts?

One of the more fun papers I've ever come across is "The Spatial Distribution of Population in 35 World Cities," by Alain Bertaud and Stephen Malpezzi, which tries to pin this down. Perhaps not very surprisingly, they find that the "density gradient" of a city flattens out, i.e., people start fleeing for the suburbs, as income and population rise—supporting Rybczynski's point that this stuff is partly inevitable and can never be totally squelched—but on the other hand, transportation costs and urban regulation do make a big difference too.

By the way, and very curiously, Bertaud and Malpezzi cite a past paper of a similar sort noting that higher crime in the central city seems to decrease sprawl while better education in the city increases it. That's a bit hard to explain, to say the least.
-- Brad Plumer 9:48 PM || ||